[PATCH] serial: amba-pl011: Enable UART in earlycon setup

[PATCH] serial: amba-pl011: Enable UART in earlycon setup

[Peter Maydell]

Currently the PL011 driver only enables the UART (by setting UARTEN
in REG_CR) in pl011_startup(), so if it is used for earlycon it is
relying on the bootrom/firmware having left the UART enabled.

There's no particular reason not to actively enable the UART before
using it for earlycon, and the earlycon handling for e.g. the 8250
UART sets up the UART in its setup function, so follow that in the
PL011.

This allows use of earlycon with a UART that the firmware hasn't
already been using for its own output, but the main motivation is
that QEMU will otherwise log a message complaining that the guest is
trying to write to a UART it never enabled.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
NB: I have tested this under QEMU, but I do not have any real
PL011-using hardware to hand to test with.
---
 drivers/tty/serial/amba-pl011.c | 35 ++++++++++++++++++++++++++++++---
 1 file changed, 32 insertions(+), 3 deletions(-)

diff --git a/drivers/tty/serial/amba-pl011.c b/drivers/tty/serial/amba-pl011.c
index 7f17d288c807..462a8c380059 100644
--- a/drivers/tty/serial/amba-pl011.c
+++ b/drivers/tty/serial/amba-pl011.c
@@ -2700,6 +2700,37 @@ static int pl011_early_read(struct console *con, char *s, unsigned int n)
  */
 static int __init pl011_early_console_setup(struct earlycon_device *device,
 					    const char *opt)
+{
+	unsigned int cr;
+
+	if (!device->port.membase)
+		return -ENODEV;
+
+	device->con->write = pl011_early_write;
+	device->con->read = pl011_early_read;
+
+	if (device->port.iotype == UPIO_MEM32)
+		cr = readl(device->port.membase + UART011_CR);
+	else
+		cr = readw(device->port.membase + UART011_CR);
+	cr &= UART011_CR_RTS | UART011_CR_DTR;
+	cr |= UART01x_CR_UARTEN | UART011_CR_RXE | UART011_CR_TXE;
+	if (device->port.iotype == UPIO_MEM32)
+		writel(cr, device->port.membase + UART011_CR);
+	else
+		writew(cr, device->port.membase + UART011_CR);
+
+	return 0;
+}
+
+OF_EARLYCON_DECLARE(pl011, "arm,pl011", pl011_early_console_setup);
+
+/*
+ * The SBSA UART has no defined control register and is assumed to
+ * be pre-enabled by firmware, so we do not write to UART011_CR.
+ */
+static int __init sbsa_uart_early_console_setup(struct earlycon_device *device,
+						const char *opt)
 {
 	if (!device->port.membase)
 		return -ENODEV;
@@ -2710,9 +2741,7 @@ static int __init pl011_early_console_setup(struct earlycon_device *device,
 	return 0;
 }
 
-OF_EARLYCON_DECLARE(pl011, "arm,pl011", pl011_early_console_setup);
-
-OF_EARLYCON_DECLARE(pl011, "arm,sbsa-uart", pl011_early_console_setup);
+OF_EARLYCON_DECLARE(pl011, "arm,sbsa-uart", sbsa_uart_early_console_setup);
 
 /*
  * On Qualcomm Datacenter Technologies QDF2400 SOCs affected by
-- 
2.43.0

Re: [PATCH] serial: amba-pl011: Enable UART in earlycon setup

[Arnd Bergmann]

On Tue, Feb 10, 2026, at 13:50, Peter Maydell wrote:
> Currently the PL011 driver only enables the UART (by setting UARTEN
> in REG_CR) in pl011_startup(), so if it is used for earlycon it is
> relying on the bootrom/firmware having left the UART enabled.
>
> There's no particular reason not to actively enable the UART before
> using it for earlycon, and the earlycon handling for e.g. the 8250
> UART sets up the UART in its setup function, so follow that in the
> PL011.
>
> This allows use of earlycon with a UART that the firmware hasn't
> already been using for its own output, but the main motivation is
> that QEMU will otherwise log a message complaining that the guest is
> trying to write to a UART it never enabled.
>
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
> NB: I have tested this under QEMU, but I do not have any real
> PL011-using hardware to hand to test with.

Looks good to me,

Acked-by: Arnd Bergmann <arnd@arndb.de>

One more question: what is the symptom without this patch with
qemu? You mentioned on IRC that qemu may warn about attempting
to write to the uart when it is disabled. Would we still get
the correct early console output before the uart is enabled?

      Arnd

Re: [PATCH] serial: amba-pl011: Enable UART in earlycon setup

[Peter Maydell]

On Tue, 10 Feb 2026 at 13:23, Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Tue, Feb 10, 2026, at 13:50, Peter Maydell wrote:
> > Currently the PL011 driver only enables the UART (by setting UARTEN
> > in REG_CR) in pl011_startup(), so if it is used for earlycon it is
> > relying on the bootrom/firmware having left the UART enabled.
> >
> > There's no particular reason not to actively enable the UART before
> > using it for earlycon, and the earlycon handling for e.g. the 8250
> > UART sets up the UART in its setup function, so follow that in the
> > PL011.
> >
> > This allows use of earlycon with a UART that the firmware hasn't
> > already been using for its own output, but the main motivation is
> > that QEMU will otherwise log a message complaining that the guest is
> > trying to write to a UART it never enabled.
> >
> > Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> > ---
> > NB: I have tested this under QEMU, but I do not have any real
> > PL011-using hardware to hand to test with.
>
> Looks good to me,
>
> Acked-by: Arnd Bergmann <arnd@arndb.de>
>
> One more question: what is the symptom without this patch with
> qemu? You mentioned on IRC that qemu may warn about attempting
> to write to the uart when it is disabled. Would we still get
> the correct early console output before the uart is enabled?

The symptom is that if you run QEMU with '-d guest_errors' to ask
it to log things the guest does that it thinks are bugs in the
guest, then you get a lot of "PL011 data written to disabled UART"
warnings. The guest still boots fine, and assuming you directed
the logging to a file rather than stdout you see the guest output.

(This is because for PL011 we have deliberately not emulated the
hardware behaviour of discarding data when the UART isn't initialized,
because we think there is a lot of "my first bare metal hello
world program" type code out there that was written against older
QEMU versions and not tested on hardware and never enables the
PL011, and we figured that making all that code break wasn't a
sensible tradeoff for theoretically increased emulation fidelity.
For newer UART models where we modelled the enable flag correctly
from the start, you would have to explicitly enable. The CMSDK
UART model is one example of that.)

I have also sent a QEMU patch to restrict the complaint to
"print this once" rather than "print this every time the
guest writes to the UART TX register".

thanks
-- PMM

Re: [PATCH] serial: amba-pl011: Enable UART in earlycon setup

[Arnd Bergmann]

On Tue, Feb 10, 2026, at 14:49, Peter Maydell wrote:
> On Tue, 10 Feb 2026 at 13:23, Arnd Bergmann <arnd@arndb.de> wrote:
>> On Tue, Feb 10, 2026, at 13:50, Peter Maydell wrote:
>>
>> One more question: what is the symptom without this patch with
>> qemu? You mentioned on IRC that qemu may warn about attempting
>> to write to the uart when it is disabled. Would we still get
>> the correct early console output before the uart is enabled?
>
> The symptom is that if you run QEMU with '-d guest_errors' to ask
> it to log things the guest does that it thinks are bugs in the
> guest, then you get a lot of "PL011 data written to disabled UART"
> warnings. The guest still boots fine, and assuming you directed
> the logging to a file rather than stdout you see the guest output.

Ok, thanks for the explanation! I still wonder whether we will
want a stable backport of the patch. It sounds like it won't
be needed for qemu (because most people won't see the harmless
warning especially after your patch) or most real hardware
(because the bootloader will have initialized the console),
but there may still be corner cases that need it.

      Arnd